Patch vastly improve latency while scanning. Slight throughput
improvements were observed as well. Is intended for improve performance
of voice and video applications, when scan is periodically requested by
user space (i.e. default NetworkManager behaviour).

Patch remove latency requirement based on PM_QOS_NETWORK_LATENCY,
this value is 2000 seconds by default (i.e. approximately 0.5 hour !?!).

Also remove listen interval requirement, which based on beaconing and
depending on BSS parameters. It can make we stay off-channel for a
second or more.

Instead try to offer the best latency that we could, i.e. be off-channel
no longer than PASSIVE channel scan time: 125 ms. That mean we will
scan two ACTIVE channels and go back to on-channel, and one PASSIVE
channel, and go back to on-channel.

Patch also decrease PASSIVE channel scan time to about 110 ms.

As drawback patch increase overall scan time. On my tests, when scanning
both 2GHz and 5GHz bands, scanning time increase from 5 seconds up to 10
seconds. Since that increase happen only when we are associated, I think
it can be acceptable. If eventually better scan time is needed for
situations when we lose signal and quickly need to decide to which AP
roam, additional scan flag or parameter can be introduced.

I tested patch by doing:

while true; do iw dev wlan0 scan; sleep 3; done > /dev/null

and

ping -i0.2 -c 1000 HOST

on remote and local machine, results are as below:

* Ping from local periodically scanning machine to AP:
Unpatched: rtt min/avg/max/mdev = 0.928/24.946/182.135/36.873 ms
Patched:   rtt min/avg/max/mdev = 0.928/19.678/150.845/33.130 ms

* Ping from remote machine to periodically scanning machine:
Unpatched: rtt min/avg/max/mdev = 1.637/120.683/709.139/164.337 ms
Patched:   rtt min/avg/max/mdev = 1.807/26.893/201.435/40.284 ms

Throughput measured by scp show following results.

* Upload to periodically scanning machine:
Unpatched: 3.9MB/s   03:15
Patched:   4.3MB/s   02:58

* Download from periodically scanning machine:
Unpatched: 5.5MB/s   02:17
Patched:   6.2MB/s   02:02
Signed-off-by: NStanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When sending authentication/association frames they
might take a bit of time to go out because we may
have to synchronise with the AP, in particular in
the case where it's really a P2P GO. In this case
the 200ms fixed timeout could potentially be too
short if the beacon interval is relatively large.

For drivers that report TX status we can do better.
Instead of starting the timeout directly, start it
only when the frame status arrives. Since then the
frame was out on the air, we can wait shorter (the
typical response time is supposed to be 30ms, wait
100ms.) Also, if the frame failed to be transmitted
try again right away instead of waiting.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

These pointers/values are never used, remove them.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

We track this, but never use it, so we can
just remove it.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Since:

commit b23b025f
Author: Ben Greear <greearb@candelatech.com>
Date:   Fri Feb 4 11:54:17 2011 -0800

    mac80211: Optimize scans on current operating channel.

we do not disable PS while going back to operational channel (on
ieee80211_scan_state_suspend) and deffer that until scan finish.
But since we are allowed to send frames, we can send a frame to AP
without PM bit set, so disable PS on AP side. Then when we switch
to off-channel (in ieee80211_scan_state_resume) we do not enable PS.
Hence we are off-channel with PS disabled, frames are not buffered
by AP.

To fix remove offchannel_ps_disable argument and always enable PS when
going off-channel and disable it when going on-channel, like it was
before.

Cc: stable@vger.kernel.org # 2.6.39+
Signed-off-by: NStanislaw Gruszka <sgruszka@redhat.com>
Tested-by: NSeth Forshee <seth.forshee@canonical.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

The probe response/beacon management frame RX code passes a
bool parameter to differentiate beacons and probe responses.
This is useless since we have the frame and can thus use its
frame control field. Moreover it is buggy since there is one
call to ieee80211_rx_bss_info with a beacon frame that is
indicated as a probe response, which is also fixed by using
the frame control field, so do that.
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When AP's SSID is hidden the BSS can appear several times in
cfg80211's BSS list: once with a zero-length SSID that comes
from the beacon, and once for each SSID from probe reponses.

Since the mac80211 stores its data in ieee80211_bss which
is embedded into cfg80211_bss, mac80211's data will be
duplicated too.

This becomes a problem when a driver needs the dtim_period
since this data exists only in the beacon's instance in
cfg80211 bss table which isn't the instance that is used
when associating.

Remove the DTIM period from the BSS table and track it
explicitly to avoid this problem.

Cc: stable@vger.kernel.org
Tested-by: NEfi Tubul <efi.tubul@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Do not scan on no-IBSS and disabled channels in IBSS mode. Doing this
can trigger Microcode errors on iwlwifi and iwlegacy drivers.

Also rename ieee80211_request_internal_scan() function since it is only
used in IBSS mode and simplify calling it from ieee80211_sta_find_ibss().

This patch should address:
https://bugzilla.redhat.com/show_bug.cgi?id=883414
https://bugzilla.kernel.org/show_bug.cgi?id=49411Reported-by: NJesse Kahtava <jesse_kahtava@f-m.fm>
Reported-by: NMikko Rapeli  <mikko.rapeli@iki.fi>
Cc: stable@vger.kernel.org
Signed-off-by: NStanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Instead of assuming 200 bytes are always enough for
all the IEs we add, give the length of the buffer
to the function and warn instead of overrunning.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Currently, mac80211 checks the DS params IE if present and
uses it for the (primary) BSS channel, instead of the one
that the frame was received on. This is particularly useful
in the 2.4 GHz band since a frame is often received on one
of the adjacent channels due to overlap.

Move this code to cfg80211 so other drivers also do this.

Additionally, on 5 GHz, in particular with some (possibly)
upcoming changes in 802.11ai and duplicate transmissions
when wider channels are used, something similar happens.
So if present, also use the (primary) channel information
contained in the HT operation IE.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

In case that there is an unsupported band, the ie will be
unallocated and the free will crash.

Cc: stable@vger.kernel.org
Signed-off-by: NDavid Spinadel <david.spinadel@intel.com>
Reviewed-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Use NL80211_SCAN_FLAG_LOW_PRIORITY flag in mac80211's scan state
machine to prematurely terminate scan operations if outbound
traffic collides. This is useful for marking background scans so
they don't affect throughput.
Signed-off-by: NSam Leffler <sleffler@chromium.org>
Signed-off-by: NAmitkumar Karwar <akarwar@marvell.com>
Signed-off-by: NBing Zhao <bzhao@marvell.com>
[set feature flag only if software scan is used]
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Instead of operating on a single channel only,
use the new channel context infrastructure in
all mac80211 code.

This enables drivers that want to use the new
channel context infrastructure to use multiple
channels, while nothing should change for all
the other drivers that don't support it.

Right now this disables both TX power settings
and spatial multiplexing powersave. Both need
to be re-enabled on a channel context basis.

Additionally, when channel contexts are used
drop the connection when channel switch is
received rather than trying to handle it. This
will have to be improved later.

[With fixes from Eliad and Emmanuel incorporated]
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Depending on the driver, channel contexts may be used or
not. If they are used, the driver must have support for
hardware scan and remain-on-channel; otherwise the driver
must not advertise support for multiple channels.

Also prohibit WDS type interfaces when channel contexts
are to be used as there's no clear definition of which
channel they use.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

removes unnecessary semicolon

Found by Coccinelle: http://coccinelle.lip6.fr/Signed-off-by: NPeter Senna Tschudin <peter.senna@gmail.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

There's no need to keep a copy of the scheduled
scan IEs after the driver has been told, if it
requires a copy it must make one. Therefore, we
can move sched_scan_ies into the function.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

In multi-channel scenarios, the channel that we will
transmit a probe request on isn't always the current
channel (which will be NULL anyway) but will instead
be the channel that the AP is on. Pass the channel
to the ieee80211_send_probe_req() function so it can
be used in the different scenarios. The scan code
continues to pass the current channel, of course.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

The optimisation of scanning only on the current
channel should check the operating channel. Also
modify it to compare channel pointer rather than
the frequency.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

ieee80211_request_sched_scan_stop() cleared
local->sched_scan_sdata. However, sched_scan_sdata
should be cleared only after the driver calls
ieee80211_sched_scan_stopped() (like with normal hw scan).

Clearing sched_scan_sdata too early caused
ieee80211_sched_scan_stopped_work to exit prematurely
without properly cleaning all the sched scan resources
and without calling cfg80211_sched_scan_stopped (so
userspace wasn't notified about sched scan completion).
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

We need to use RCU to assign scan_sdata.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Some drivers (iwlegacy, iwlwifi and rt2x00) today use the
bss_conf.last_tsf value. By itself though that value is
completely worthless since it may be ancient. What really
is needed is synchronisation between some device time and
the TSF.

To clarify this, rename bss_conf.last_tsf to sync_tsf and
add sync_device_ts which is obtained from rx_status which
gets a new field device_timestamp for this purpose. This
is intentionally not using the mactime field since that
is used for other things and in IBSS is expected to sync
with the IBSS's TSF which isn't necessarily true for the
device timestamp.

Also, since we have the information and it's useful even
before the connection has been established, give all the
timing details to the driver before authenticating.
Reviewed-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Scan receive is rather inefficient when there are
multiple virtual interfaces. We iterate all of the
virtual interfaces and then notify cfg80211 about
each beacon many times.

Redesign scan RX to happen before everything else.
Then we can also get rid of IEEE80211_RX_IN_SCAN
since we don't have to accept frames into the RX
handlers for scanning or scheduled scanning any
more. Overall, this simplifies the code.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Instead of tracking whether or not we're in a
scheduled scan, track the virtual interface
(sdata) in an RCU-protected pointer to make it
usable from RX to check the MAC address.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Making the scan_sdata pointer usable with RCU makes
it possible to dereference it in the RX path to see
if a received frame actually matches the interface
that is scanning. This is just preparations, making
the pointer __rcu.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

The function building probe-request IEs does not validate the band is
supported before dereferencing it. This can result in a panic when
all bands are traversed, as done during sched-scan start.

Warn when this happens and return an empty probe request. Also fix
sched-scan to not waste memory on unsupported bands.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Redesign all the off-channel code, getting rid of
the generic off-channel work concept, replacing
it with a simple remain-on-channel list.

This fixes a number of small issues with the ROC
implementation:
 * offloaded remain-on-channel couldn't be queued,
   now we can queue it as well, if needed
 * in iwlwifi (the only user) offloaded ROC is
   mutually exclusive with scanning, use the new
   queue to handle that case -- I expect that it
   will later depend on a HW flag

The bigger issue though is that there's a bad bug
in the current implementation: if we get a mgmt
TX request while HW roc is active, and this new
request has a wait time, we actually schedule a
software ROC instead since we can't guarantee the
existing offloaded ROC will still be that long.
To fix this, the queuing mechanism was needed.

The queuing mechanism for offloaded ROC isn't yet
optimal, ideally we should add API to have the HW
extend the ROC if needed. We could add that later
but for now use a software implementation.

Overall, this unifies the behaviour between the
offloaded and software-implemented case as much
as possible.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Adding casts of objects to the same type is unnecessary
and confusing for a human reader.

For example, this cast:

	int y;
	int *p = (int *)&y;

I used the coccinelle script below to find and remove these
unnecessary casts.  I manually removed the conversions this
script produces of casts with __force and __user.

@@
type T;
T *p;
@@

-	(T *)p
+	p
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use the new bool function ether_addr_equal to add
some clarity and reduce the likelihood for misuse
of compare_ether_addr for sorting.

Done via cocci script:

$ cat compare_ether_addr.cocci
@@
expression a,b;
@@
-	!compare_ether_addr(a, b)
+	ether_addr_equal(a, b)

@@
expression a,b;
@@
-	compare_ether_addr(a, b)
+	!ether_addr_equal(a, b)

@@
expression a,b;
@@
-	!ether_addr_equal(a, b) == 0
+	ether_addr_equal(a, b)

@@
expression a,b;
@@
-	!ether_addr_equal(a, b) != 0
+	!ether_addr_equal(a, b)

@@
expression a,b;
@@
-	ether_addr_equal(a, b) == 0
+	!ether_addr_equal(a, b)

@@
expression a,b;
@@
-	ether_addr_equal(a, b) != 0
+	ether_addr_equal(a, b)

@@
expression a,b;
@@
-	!!ether_addr_equal(a, b)
+	ether_addr_equal(a, b)
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This based on an idea posted by Stanislaw Gruszka,
though I accept full blame for the implementation!

This has been tested with ath9k.

The idea is to let users scan on the current operating
channel without interrupting normal traffic more than
absolutely necessary (changing power level might reset
some hardware, for instance).
Signed-off-by: NBen Greear <greearb@candelatech.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

its not used where, while we directly obtain ieee80211_bss's
pointer in ibss.c by calling cfg80211_get_bss
Signed-off-by: NMohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Before we send probes in connection monitoring we check if scan is not
pending. But we do that check without locking. Fix that and also do not
start scan if connection monitoring is in progress.
Signed-off-by: NStanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Whenever the station informs the AP that it is about to leave the
operating channel, the timestamp should be recorded. It is handled
in scan resume but not in scan start. Fix that.
Signed-off-by: NRajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

mac80211 is lenient with respect to reception of corrupted beacons.
Even if the frame is corrupted as a whole, the available IE elements
are still passed back and accepted, sometimes replacing legitimate
data.  It is unknown to what extent this "feature" is made use of,
but it is clear that in some cases, this is detrimental.  One such
case is reported in http://crosbug.com/26832 where an AP corrupts
its beacons but not its probe responses.

One approach would be to completely reject frames with invaid data
(for example, if the last tag extends beyond the end of the enclosing
PDU).  The enclosed approach is much more conservative: we simply
prevent later IEs from overwriting the state from previous ones.
This approach hopes that there might be some salient data in the
IE stream before the corruption, and seeks to at least prevent that
data from being overwritten.  This approach will fix the case above.

Further, we flag element structures that contain data we think might
be corrupted, so that as we fill the mac80211 BSS structure, we try
not to replace data from an un-corrupted probe response with that
of a corrupted beacon, for example.

Short of any statistics gathering in the various forms of AP breakage,
it's not possible to ascertain the side effects of more stringent
discarding of data.
Signed-off-by: NPaul Stewart <pstew@chromium.org>
Cc: Sam Leffler <sleffler@chromium.org>
Cc: Eliad Peller <eliad@wizery.com>
Acked-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Because of the constant size and guaranteed 16 bit alignment, the inline
compare_ether_addr function is much cheaper than calling memcmp.
Signed-off-by: NFelix Fietkau <nbd@openwrt.org>
Acked-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

when we run high bandwidth UDP traffic and we trigger a scan, the scan
state machine seems to be looping in SUSPEND->RESUME->DECISION->SUSPEND
and SET_CHANNEL seems to be never called as 'tx_empty' is never true
while running UDP traffic. fix this by settting SET_CHANNEL state when
we get into RESUME state.

Cc: Leela Kella <leela@qca.qualcomm.com>
Signed-off-by: NMohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

DaveM said:
   Please, this kind of stuff rots forever and not using bool properly
   drives me crazy.

Joe Perches <joe@perches.com> gave me the spatch script:

	@@
	bool b;
	@@
	-b = 0
	+b = false
	@@
	bool b;
	@@
	-b = 1
	+b = true

I merely installed coccinelle, read the documentation and took credit.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The on-channel work optimisations have caused a
number of issues, and the code is unfortunately
very complex and almost impossible to follow.
Instead of attempting to put in more workarounds
let's just remove those optimisations, we can
work on them again later, after we change the
whole auth/assoc design.

This should fix rate_control_send_low() warnings,
see RH bug 731365.

Cc: stable@vger.kernel.org
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Attempting to micro-optimise the scan by going
fully live again when scanning the operating
channel just made the code extremely complex
and has little gain in most use cases. Remove
all that code and simplify the state machine
again.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

These files are non modular, but need to export symbols using
the macros now living in export.h -- call out the include so
that things won't break when we remove the implicit presence
of module.h from everywhere.
Signed-off-by: NPaul Gortmaker <paul.gortmaker@windriver.com>